goblossom links international artists to countries, official accounts, music releases, source coverage, career history, and dated audience observations.
| Record | Key fields and relationships |
|---|---|
| Countries & artist-country associations | Country code, names, target count, artist ID, citizenship/origin association, candidate position, ranking basis, evidence URL, checked date. Artists can belong to several rosters. |
| Artists | Stable Wikidata ID, name, native name, description, genres, languages, source-reported website, source provenance. One identity across countries. |
| Artist accounts | Platform, account ID, URL, artist ID, verification status, evidence, checked date. Source-reported is distinct from first-party verified. |
| Metric snapshots | Artist, platform, metric, account ID, exact value, observation date and date precision, method, evidence URL. Append-only time series; compare only matching platform, metric, and account. |
| Labels & relationships | Artist, label/manager/distributor, role, current/historical/unconfirmed status, start/end dates where documented, evidence URL, checked date. |
| Career events | Artist, event type, date with source precision, title, explanation, source URL. Missing dates are not invented. |
| Releases | Artist, title, format, release date, language, official announcement/listening URL, editorial notes. Imported release lists are partial. |
| Coverage & captures | Artist, source URL, publisher, original language, date, original summary, source text/notes, content fingerprint. Capture deduplication uses normalized URLs or text hashes. |
| Media assets | Story, image URL, original source URL, alt text, creator credit, license. Discovery of an image does not grant reuse rights. |
| Stories & revisions | Artist/release/source relationships, headline, dek, body, format, SEO fields, social caption, newsletter blurb, verification checks, immutable revision snapshots. |
| Generation jobs | Idempotency key, input snapshot, generator/version, output story, status, errors and timestamps. The active generator is a structured factual release-brief template. |
| Publications | Story, exact reviewed revision, publishing actor/time, withdrawal time. Publication requires saved fact and rights checks. |
| Monitoring sources | Artist, URL, platform, retrieval method, status, last check/error. On-demand historical Wikidata metric refresh is available; no recurring scheduler or authenticated platform collection is active. |
The initial country inventory uses Wikidata ISO country codes and includes territories and additional music markets. Country candidates are gathered from citizenship or country-of-origin relationships using MusicBrainz identifiers and supplementary musician-occupation queries, then filtered for music-related occupations or group types. The first 20 qualifying candidates by sitelink count form the provisional roster. A broad legacy catalog is useful for history; current-popularity rosters require a separate dated ranking built from platform or chart evidence.
The initial account and label data is sourced from Wikidata statements. An undated label relationship is unconfirmed, not a claim about a current contract. Language fields describe source-recorded languages, not necessarily every language the artist records in. Individual profiles link to the source record.
Browse and search country rosters; inspect sourced artist profiles; add releases, coverage, captured updates, metric snapshots, label relationships, and career events; refresh historical metric observations on demand; create factual release briefs with social/newsletter derivatives; revise, review, publish, and withdraw stories.
Continuous discovery, recurring social monitoring, authenticated streaming/social APIs, automatic translation, and model-generated reviews are not active. No invented follower counts, listening experiences, quotes, or unverified image permissions are used to fill gaps. A complete, current Top 20 per country still needs local-source validation and current popularity evidence.
Download the SQLite database, CSV tables, and coverage report. The imported catalog is versioned with the application. Editorial additions and metric histories persist in a relational database. Imported artists are materialized into the database on first write. This preserves a reproducible source baseline while allowing ongoing research.
Download the catalog JSON · Download the artist index CSV · Download country coverage CSV · Download SQL schema
Baseline structured data: Wikidata CC0. Linked articles, social pages, and images retain their own rights.
Open the verification desk. The original import stays immutable. A reviewed overlay records corrections, aliases, before/after values, claim-level evidence, status, reviewer, and check time. Website observations record requested URL, final URL, HTTP outcome, page title, content hash, and exact outgoing-link matches. Raw article bodies are not republished.
Account states: source_reported; corroborated_link_review_pending; first_party_link_verified. The last state verifies a reviewed link from an artist site, not authenticated account ownership or audience statistics. A page fetch, a matching name, or an encyclopedia entry never verifies an entire profile.
Verification is versioned alongside the catalog and survives reassembly using the reviewed overlay. These evidence records are not a new editable database table; editorial additions retain their existing relational storage. The research-queue CSV describes the original baseline; corrected values and subsequent evidence are in the corrected catalog and verification export.
Ranking publication requires an explicit country-association definition, eligible candidate population, observation window, comparable metrics, source dates, territorial scope, missing-data treatment, and editorial approval. No current Top 20 ranking is verified by this pass. Missing metrics stay unavailable, not zero; record-label website presence does not establish contractual exclusivity.
Download evidence and decisions · Download baseline research queue